Complex Seismic Wavefield of Long-Period Components around the Nankai Trough in Southwest Japan and its Effect on Source Mechanism Analyses

Abstract:

Previous simulation studies on seismic waves for a suboceanic earthquake showed that the long-period component can be significantly amplified and prolonged during their propagation to land areas: e.g., Furumura et al. (2008) indicated that low-velocity sediment layers such as accretionary prisms around the trough area considerably affected the propagation of surface waves and were the cause of the long-period motions. Such investigations on the motions and their relations with subsurface structures are very important in discussing disaster mitigation for future subduction-zone earthquakes. In this study, we simulated seismic wavefields for small- to moderate-sized suboceanic earthquakes and investigated the propagation of surface waves at each period of long-period ranges in land and seafloor areas.

Our simulation results presented interesting features on the propagation pattern in the areas. The Rayleigh wave at a central period of 10 s propagates with a slower speed in seafloor area than that in land area because of their strong dependence on shallow structures with low seismic velocities such as sediment layers, an oceanic layer, and oceanic crusts. For a period of 15 s, however, the Rayleigh wave propagates with a faster speed in seafloor area than that in land area because of the dependence to oceanic mantle with high velocities. For a period of 20 s, the differences of the propagation between land and seafloor areas are not clearly found. These complex propagation patterns are also demonstrated in the spatial variations of phase and group velocities for the fundamental mode of Rayleigh waves and are explained by velocity differences in the structures between land and ocean areas. Our results indicate that either the point-source or the finite-source waveform analyses in the long-period range using a combined land and seafloor data would potentially produce incorrect evaluations and results if the velocity differences are not correctly incorporated into the analyses.